According to Jim Brander:

gpc-typeck.c:626

...

626 for (r =3D rhs; r; r =3D TYPE_LANG_SPECIFIC (r)->elts [1]) (gdb) l 621 tree lhs, rhs; 622 { 623 register tree r; 624 if (TREE_CODE (lhs) =3D=3D RECORD_TYPE 625 && TREE_CODE (rhs) =3D=3D RECORD_TYPE) 626 for (r =3D rhs; r; r =3D TYPE_LANG_SPECIFIC (r)->elts [1]) 627 { 628 if (comptypes (lhs, r)) 629 return 1; 630 }

Please try the following patch: replace

if (TREE_CODE (lhs) == RECORD_TYPE && TREE_CODE (rhs) == RECORD_TYPE)

by

if (IS_OBJECT_TYPE (lhs) && IS_OBJECT_TYPE (rhs))

and let us know the result.

You discovered a severe bug. Thanks.

Peter

e-mail: peter.gerwinski@uni-essen.de home address: D"usseldorfer Str. 35, 45145 Essen, Germany WWW: http://agnes.dida.physik.uni-essen.de/~peter/